STUDY PROTOCOL Open Access

Mentalization-based treatment in groupsfor adolescents with borderline personalitydisorder (BPD) or subthreshold BPD versustreatment as usual (M-GAB): study protocolfor a randomized controlled trialEmma Beck1,2,3* , Sune Bo1,2, Matthias Gondan3, Stig Poulsen3, Liselotte Pedersen2,3, Jesper Pedersen1

and Erik Simonsen2,4

Abstract

Background: Evidence-based outpatient psychotherapeutic programs are first-line treatment of borderline personalitydisorder (BPD). Early and effective treatment of BPD is crucial to the prevention of its individual, psychosocial, andeconomic consequences. However, in spite of recent advantages in diagnosing adolescent BPD, there is a lack ofcost-effective evidence-based treatment programs for adolescents. Mentalization-based treatment is an evidence-basedprogram for BPD, originally developed for adults.

Methods/Design: Aims/hypotheses: We will investigate whether a specifically designed mentalization-based treatmentin groups is an efficacious treatment for adolescents with BPD or subthreshold BPD compared to treatment as usual.The trial is a four-center, two-armed, parallel-group, assessor-blinded randomized clinical superiority trial. One hundredtwelve patients aged 14 to 17 referred to Child and Adolescent Psychiatric Clinics in Region Zealand are randomized to1 year of either mentalization-based treatment in groups or treatment as usual. Patients will be included if they meet atleast four DSM-5 criteria for BPD. The primary outcome is self-reported borderline features at discharge. Secondaryoutcomes will include self-harm, depression, BPD criteria, externalizing and internalizing symptoms, and socialfunctioning, together with parental reports on borderline features, externalizing and internalizing symptoms. Measures ofattachment and mentalization will be included as mediational variables. Follow-up assessment will take place at 3 and12 months after end of treatment.

Discussion: This is the first randomized controlled trial to test the efficacy of a group-based mentalization-basedtreatment for adolescents with BPD or subthreshold BPD. If the results confirm our hypothesis, this trial will addto the treatment options of cost-effective treatment of adolescent BPD.

Trial registration: Clinicaltrials.gov NCT02068326, February 19, 2014.

Keywords: Mentalization-based treatment, Adolescence, Borderline personality disorder, Group psychotherapy,Mentalizing

* Correspondence: ebk@regionsjaelland.dk1Child and Adolescent Psychiatric Department, Region Zealand, Smedegade16, 4000 Roskilde, Denmark2Psychiatric Research Unit, Region Zealand, Fælledvej 6, 4200 Slagelse,DenmarkFull list of author information is available at the end of the article

© 2016 Beck et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Beck et al. Trials (2016) 17:314 DOI 10.1186/s13063-016-1431-0

BackgroundBorderline personality disorder (BPD) is a severe mentaldisorder with symptoms such as affective instability, impul-sivity and self-harm [1]. In comparison to other personalitydisorders (PDs), BPD is associated with more severe impair-ments in social functioning [2, 3], higher rates of psychiatriccomorbidity and increased risk of suicide attempts [4].Prevalence in the general population ranges from 0.5 to2.7 % [5–9]. In clinical populations the prevalence rangesbetween 9 and 22 % in outpatient settings [10, 11], and upto 40 % in inpatient settings [12]. Research into the etiologyof BPD indicates that it is multifactorial and includes gen-etic dispositions, neuropsychological dysfunctions, and en-vironmental factors such as neglect and physical abuse [13].It has been disputed whether it is possible and mean-

ingful to diagnose PDs in general and BPD in particularin adolescents [14–16]. However, there is evidence thatBPD can be diagnosed reliably and validly in adolescentswith reliability rates comparable to those found withadults [17–22]. The possibility of diagnosing adolescentPD has been introduced in the Diagnostic and StatisticalManual of Mental Disorders, fifth edition (DSM-5) [1]and BPD prevalence rates among adolescent inpatientsare similar to those of adult inpatients [23, 24]. Stabilityof PD diagnoses among adolescents also resembles thatof adults [25].BPD is associated with both chronic medical illnesses

and comorbid psychiatric disorders, and psychiatric co-morbidity may be even higher in adolescents than inadults [26–31]. Patients with BPD tend to use inpatientand outpatient treatment far more than patients withother PDs [32] leading BPD to be associated with costlyhealth service use [27, 33], also in adolescence [34, 35].Therefore, the development of early and effective treat-ment programs, which include adolescents with BPD atthe threshold level is important and may have long-termbenefits for patients, their families and society [36–40].While pharmacological treatments can reduce specificsymptoms experienced by BPD patients, they fail tobring about a fundamental or long-lasting change in thedisorder itself. In contrast, several structured outpatientpsychotherapeutic programs such as dialectical behaviortherapy (DBT), mentalization-based treatment (MBT),transference-focused therapy (TFP), cognitive behavioraltherapy (CBT) and schema-focused group therapy (SFT)have been shown to be efficacious. Although further evi-dence is still warranted, psychotherapy is suggested asthe primary treatment for BPD [41, 42].Adaptation of structured and manualized treatment

programs to adolescents with BPD is novel, and relativelyfew controlled studies have tested their efficacy: In an ran-domized controlled trial (RCT) by Mehlum et al., DBT-Awas found to be superior to the control condition in redu-cing the frequency of self-harm, both at treatment end

and at follow-up [43, 44]. Other studies on feasibility andefficacy of DBT-A also indicate a positive outcome forDBT-A [45–49], but these studies are uncontrolled, smallscale, and use different outcome measures, leading theevidence base to be considered insufficient [50]. As thetarget population for DBT-A is adolescents with suicidaland self-harming behavior, the overlap between thisgroup and adolescents with BPD is only partial [50],and the efficacy of DBT-A for treating BPD in adoles-cents is unknown. Chanen et al. [51] tested individualcognitive analytic therapy (CAT) for adolescents withBPD symptoms and found no significant differences be-tween the outcomes of CAT and the control conditionat 24 months, but with faster rates of symptomatic im-provement in the CAT condition. Schuppert et al. [52]also found no superiority of a 17-session group-basedemotion regulation training course. In contrast, Rossouwand Fonagy’s [53] adaptation of mentalization-based treat-ment (MBT) for borderline personality patients to adoles-cents yielded more promising results: A 1-year treatmentprogram for adolescents with self-harming behavior (ofwhom 73 % met the diagnostic criteria for BPD) compris-ing weekly individual MBT sessions and monthly family-therapy sessions, was more effective than treatment asusual (TAU). In an uncontrolled study, Laurenssen et al.[54] adapted MBT to an inpatient setting with adolescentBPD patients, and found a significant improvement inAxis-I symptomatology and personality functioning.MBT is specifically developed to treat BPD and its effi-

cacy of treating BPD in adults has been tested in severalRCTs [55–58]. MBT is based on attachment theory andpsychodynamic principles and has a high degree of struc-ture and a clear treatment goal of improving patients’mentalizing skills [55–59]. Mentalization has been definedas the capacity to understand and interpret - implicitlyand explicitly - one’s own and others’ behavior as anexpression of mental states such as feelings, thoughts, fan-tasies, beliefs, and desires [60]. The capacity develops dur-ing childhood and is intimately linked with the quality ofearly attachment relationships [61]. When the caregiversaffective mirroring of the child’s mental state is markedand congruent, i.e., is representative of the child’s mentalstate, and not her own, it serves as a representation of thestate that the child can incorporate into its own represen-tation of self [60]. The theoretical framework of MBTlinks BPD pathology to an impairment in the capacity tomentalize. In states of emotional arousal, BPD patients arevulnerable to shifting toward a non-mentalizing mode. Inshort, MBT offers therapeutic techniques for identifyingwhen this shift is occurring and how to help the patientreturn to a mentalizing mode, where an understanding ofone’s own and others’ minds can be used to regulate thepatient’s emotional state [40]. In the original model forMBT, group psychotherapy is supported by individual

Beck et al. Trials (2016) 17:314 Page 2 of 13

psychotherapy that ensures patients’ attendance to groupsessions by motivating and working constructively withpatients’ negative experiences that could otherwise leadthem to drop out of treatment [59]. However, such a rela-tively resource-demanding treatment model can impedeimplementation, and research into the effectiveness of theseparate modalities of MBT is sparse.While the overall aim of developing patients’ mentaliz-

ing skills overlaps in individual and group therapy, theremay be some particular advantages to using the groupmodality in the treatment of BPD. Karterud [62] summa-rizes the literature on these advantages as follows: First,the relatively intimate two-person relationship of indi-vidual therapy is likely to activate attachment patternsand transference/countertransference that are emotion-ally too burdening for the BPD patient to endure. In thegroup, such interpersonal processes will be “spread out”on different group members and therefore be experi-enced as less intense. Second, as the interpersonal diffi-culties associated with BPD naturally unfold betweengroup members, group therapy offers an opportunity toexplore and work on them in vivo. Third, difficulty withauthority figures experienced by many BPD patients candiminish their receptiveness to feedback coming fromthe therapist. In contrast, when feedback comes fromgroup members, this difficulty may be bypassed. Further-more, compared to individual therapy, group therapy isconsiderably less expensive.Including parents in the MBT program for adolescent

BPD and targeting the parent-child relationship is im-portant. Parental and family circumstances are contrib-uting factors to the development of BPD [63, 64], andBPD symptoms such as affect dysregulation and impul-sivity are especially prone to be exhibited within thecontext of attachment relationships [40], such as the re-lationship between adolescents with BPD and their par-ents. We suggest that guidance for parents on how tounderstand and respond to their child in interpersonalstressful situations can promote mentalization and co-regulation within the parent-adolescent relationship, andthat these intersubjective experiences may serve as apathway to development of the adolescents’ mentalizingskills. We also propose that it will reduce prematuredropout to educate and guide the parents about the im-portance of supporting their child in times of low motiv-ation for attending treatment sessions [65].To our knowledge, group-based MBT for adolescent

BPD or subthreshold BPD that includes parents in thetreatment has not yet been tested for efficacy in a con-trolled study. However, in an uncontrolled feasibility trial(N = 25) of a group-based MBT program similar to theone presented in the present protocol, we found symp-tomatic improvement in 92 % of the patients. On the Bor-derline Personality Features Scale for Children (BPFS-C),

which is also the primary outcome measure in thepresent trial, the difference in before and after scoreswere highly significant (Bo, Sharp, Beck, Pedersen,Gondan and Simonsen (submitted)).

Aims and hypothesesWe will investigate whether a specifically designed treat-ment program, Mentalization-based treatment in groupsfor adolescents with BPD or subthreshold BPD, is anefficacious treatment compared to treatment as usual(TAU). The MBT program includes an introduction tomentalization (MBT-I), mentalization-based group ther-apy (MBT-G) and mentalization-based psychoeducationfor the parents of the patients (MBT-P) [62, 66]. Thepresent study will test if this specifically designed group-based MBT program is superior to TAU as measured bya decrease in borderline personality features after thelast MBT-G session (session no. 40).

Methods/DesignThe study is a randomized two-armed, parallel group,assessor-blinded continuous outcome superiority trial,comparing a group-based MBT program with TAU in112 adolescents with BPD or subthreshold BPD, follow-ing intention-to-treat (ITT) principles. The study will becarried out at four outpatient child and adolescent psy-chiatric clinics within The Child and Adolescent Psychi-atric Department, Region Zealand, Denmark.

Eligibility criteriaInclusion criteria

1. Age from 14 to 17 years2. Meeting a minimum of four DSM-5 BPD criteria3. Total score higher than clinical cutoff (67) on the

Borderline Personality Features Scale for Children(BPFS-C)

4. Parents’ or parent substitutes’ commitment toparticipate in the MBT-P program and to supporttheir child’s participation in the MBT-I and MBT-Gprogram

5. Written informed consent

Exclusion criteria

1. Comorbid diagnosis of pervasive developmentaldisorder, learning disability (IQ < 75), anorexia,current psychosis, diagnosis of schizophrenia orschizotypal personality disorder and antisocialpersonality disorder (DSM-5)

2. Any other mental disorder other than BPD consideredthe primary diagnosis

3. Current (past 2 months) substance dependence (butnot substance abuse)

Beck et al. Trials (2016) 17:314 Page 3 of 13

4. Current psychiatric inpatient treatment5. Receiving any other psychotherapeutic treatment6. Not living with parent(s) or parent’s substitute(s),

who are able to participate in the MBT-P program7. Not fluent in Danish

Recruitment and baseline proceduresParticipants from child and adolescent psychiatric in-and outpatient clinics in Region Zealand will be screenedfor eligibility as part of their psychiatric review and re-ferred to further assessment, provided they do not meetany exclusion criteria. A simple screening instrumenthas been developed, and staff will be trained in using itprior to recruitment. When clinicians encounter patientswho fulfil the inclusion criteria, the patients are sent toan assessor, who will inform the patient and parentsthoroughly about participation in the M-GAB trial in-cluding the assessment procedure. Patients are then in-vited to two sessions of assessment for eligibility. Bothsessions last between 2 and 3 hours including breaks.Subsequently the family is told if they are invited to par-ticipate in the randomization, and if so, outcome mea-sures at baseline are administered. Randomization willtake place after completion of the baseline assessmentand patients will be allocated to either the MBT ingroups or the TAU group (see below). The flowchart andmeasurement points are presented in Fig. 1.

InterventionsBoth treatment conditions, MBT in groups and TAU,will be delivered at all four child and adolescent psychi-atric clinics in Region Zealand.

Experimental intervention: MBT in groupsMentalization-based group therapy is constituted by a setof treatment principles. Naturally, some principles over-laps with individual MBT (i.e., “regulating arousal” and“keeping an affect focus”), but others are specific to thegroup component (i.e., identifying and mentalizing eventsin the group) [62, 67]. The patients’ profound difficultieswith interpersonal functioning and affect regulation re-quire therapeutic techniques potent enough to counteractthe difficulties that can arise during sessions. A maintherapeutic goal is to balance the management of therapistauthority and structure of each session with the open-minded, exploring, and curious stance in order “to createand sustain a mentalizing discourse in the group” [62, 67].MBT in groups is a 1-year psychotherapy program

with three components, MBT-Introduction (MBT-I),MBT-Group (MBT-G) and MBT-Parents (MBT-P).MBT-I is a structured 3-week introductory psychoedu-

cative program for patients covering the concepts of men-talizing and attachment, and information about BPD.MBT-I includes role plays and discussions of cases and

video clips in order to encourage active participation.Furthermore, the group members will be encouraged toparticipate in small homework assignments between thesessions. MBT-I was developed by Karterud and Bateman[66], and modified for our purpose in collaboration withthe authors.Mentalization-based group therapy (MBT-G) consists

of 37 weekly sessions of mentalization-based psychother-apy in groups. Sessions are not accompanied by continu-ing individual sessions. However, case formulations aremade with patients during individual sessions with boththerapists, and this process scaffolds and supports thetherapeutic work in the group. The overall purpose ofthe case formulation sessions is for therapists and patientto develop a clear mutual understanding of the patient’smain difficulties and psychotherapeutic focus points, whichcan be drawn on during group sessions. The case formula-tion is to be formulated within the mentalization-basedframework, applying the theoretical understandings andprinciples from mentalization theory. The individual focusin these sessions organizes thinking for therapists and pa-tient and supports the mentalizing work in the ensuinggroup sessions [62]. Three individual case formulation ses-sions are scheduled before the initiation of the MBT-G.The patient will also be informed which therapist is his orher contact person, and an updated crisis plan with thetherapist’s contact details is included in the formulation.After eight to ten group sessions and again toward the endof the program (after 25 group sessions), a case formula-tion session is scheduled for revising the formulationaccording to the deepened knowledge and development ofthe patient.MBT-P is a psychoeducation program for the patients’

parents running parallel to MBT-I and MBT-G. It is aslow-open six-session program comprising role plays,plenary analysis of difficult interpersonal events betweenparent and adolescent and therapists’ presentations withinformation on BPD, mentalization, and attachment.The importance of parents’ continuous support of theiradolescents’ attendance to treatment is also discussed.All sessions have a length of 90 minutes. MBT-P is adaptedfrom Karterud and Bateman’s manual [66]. An overview ofthe entire MBT in groups program is presented in Fig. 1.Both MBT-P as well as the contact-person function men-tioned above are intended to prevent premature dropoutof the treatment.

TherapistsMBT will be delivered by clinical psychologists and apsychiatrist. Prior to starting the MBT treatment, thera-pists will have participated in a 2-day introduction toMBT theory and basic principles and a 5-day trainingprogram by Professor Sigmund Karterud, who developedthe manual for MBT-G and MBT-I in collaboration with

Beck et al. Trials (2016) 17:314 Page 4 of 13

Anthony Bateman [62, 66]. Replacement of therapistsduring the trial may be necessary, in which case at leastone therapist in each group will have finished training,while the co-therapist is undertaking or waiting to begintraining. National specialists in MBT and adolescentpsychiatry provide 1–2 hours of supervision per monthto secure adherence to the manual. All sessions arevideotaped and 10 % are randomly selected for ratings ofadherence to the treatment manual applying the MBT-Gadherence and competence scale [62]. Therapists deliv-ering MBT within this trial will not deliver TAU, butmay carry out non-MBT-based clinical work outside thetrial such as clinical assessments or case managementfor in-patients.

Organization of MBT in groups within the four clinicsWhenever patients terminate/drop out of a MBT-Ggroup and new patients can be admitted into the group,

a new MBT-I group starts up. As a default, patients willbe offered MBT-I in their local clinic, but may alsochoose to enter MBT-I in a different clinic in order toparticipate in a larger group. MBT-I groups starting upwith less than five patients are run by one therapist.Groups with five or more patients are delivered by twotherapists, Patients start MBT-G treatment immediatelyafter completion of MBT-I. The MBT-G treatment willrun in five slow-open groups across the four clinics witha maximum of eight patients in each.

Treatment as usualTAU is based on supportive techniques and comprisespsychoeducation, counseling and, if needed, ad hoc crisismanagement with the overall aim of alleviating BPDsymptomatology. TAU is delivered by child and adoles-cent mental health professionals with thorough experi-ence in working with psychopathology in adolescents.

Fig. 1 Flow diagram of M-GAB study design. Consort flow diagram

Beck et al. Trials (2016) 17:314 Page 5 of 13

They will be psychologists, psychiatrists, social workersor nurses employed in child and adolescent psychiatricclinics in Region Zealand. Adherence to the supportiveapproach is monitored in regular supervision, as TAU isnot manualized. For the purpose of this study, we defineda limit for the minimum treatment by standardizing to atleast 12 individual monthly sessions. Additional contactmay vary across clinics and according to the needs of thepatient. The supervisors and staff selected to carry out theTAU treatment and supervision have no previous MBTtraining. They will not be undertaking MBT training dur-ing the trial, nor will they deliver MBT to patients in- oroutside the trial. We will monitor and register both theMBT and TAU patients’ contact to the health system dur-ing the treatment period, to be able to report both quanti-tatively and qualitatively the treatment both groups havereceived.

Medication in the experimental intervention group and inTAUAll patients will continuously be monitored for their useof medication during the psychotherapeutic interven-tion, and any differences between the groups will be ana-lyzed as a possible confounder for outcome.A protocol containing guidelines for pharmacological

treatment is provided (available on request) and medica-tion will be registered. The protocol will follow thenational recommendations for treating mental disordersin adolescents, and more specifically the new guidelinesreleased in June 2015 from the National Board of Healthfor treatment of borderline personality [68]. All psychia-trists responsible for medical treatment will be commit-ted to this medical protocol adherence.

Assessment and instrumentsDiagnostic evaluation at baselineBaseline assessment will be carried out prior torandomization by the first author and a research assist-ant, both of whom are clinical psychologists. The firstauthor will not carry out treatment in any of the treat-ment arms. The research assistant may deliver MBT-I tosmaller groups of patients. He or she will not deliverTAU.Clinical syndromes: general psychopathology will be

assessed with the Mini-International NeuropsychiatricInterview for children and adolescents (MINI-KID 6.0)[69], which is a structured diagnostic interview for chil-dren aged between 6 and 17 years old based on 24DSM-IV and ICD-10 criteria. The MINI-KID generatesreliable and valid psychiatric diagnoses for children andadolescents [70].Personality disorders will be assessed with the Struc-

tured Clinical Interview for DSM-IV-Axis II (SCID-II)[71]. SCID-II assesses the presence of personality

disorders listed in the DSM-5. The SCID-II is consideredthe “gold-standard” assessment instrument for personal-ity disorders [72], and has shown good psychometricproperties [71]. All SCID-II interviews will be carriedout by the first author who is a trained and experiencedSCID-II interviewer.Borderline personality disorder. Zanarini et al.’s [73]

Childhood Interview for DSM-IV Borderline PersonalityDisorder (CI-BPD) was developed to also allow the iden-tification of subthreshold borderline pathology in chil-dren. The CI-BPD is a semi-structured interview inwhich the interviewer asks a series of questions and sub-sequently rates each DSM-based criterion on a score of0 (absent), 1 (probably present) or 2 (definitely present).The CI-BPD has excellent psychometric properties [73, 74]and has demonstrated significant, albeit moderate, agree-ment to clinician diagnosis at time of discharge in a sampleof inpatients (kappa = .47) and good internal consistencywith a Cronbach’s alpha of 0.82 [75]. The first author hasbeen trained in the CI-BPD by Zanarini in person and willtrain all other assessors in the study to secure good inter-rater reliability.

OutcomesPrimary outcomeThe primary outcome, the Borderline Personality FeaturesScale for Children (BPFS-C) [76], is a 24-item dimensionalmeasure adapted from the borderline scale of the Person-ality Assessment Inventory (PAI) [77] for use with chil-dren and adolescents from 9 years of age. It has the samefour subscales as the PAI borderline scale: Negative Rela-tionships, Affective Instability, Self-harm and IdentityProblems. The 24 items are rated on a 5-point Likertscale, ranging from 1 (not at all true) to 5 (always true)with a higher total score indicating more severe levels ofborderline personality features. Crick et al. [76] establishedevidence for the construct validity and demonstrated highinternal consistency. Evidence for cross-informant con-cordance, criterion and concurrent validity were estab-lished by Sharp and colleagues [78]. Satisfactory constructand criterion validity has also been found for the abbrevi-ated 11-item version of the BPFS-C [79]. In an RCT of in-dividual MBT for adolescents [53], the BPFS-C was foundto be sensitive to clinical change.

Secondary outcome measuresDepression will be measured with the Beck’s DepressionInventory for Youth (BDI-Y) from Becks Youth Inven-tories of emotional and social impairment, which havealso proven valid and reliable in Danish [80].Self-harm will be measured using the self-report scale

on self-harm from the Risk-Taking and Self-Harm Inven-tory for adolescents (RTSHIA) [81]. The RTSHIA is a38-item measure adapted from the adult Self-Harm

Beck et al. Trials (2016) 17:314 Page 6 of 13

Inventory [82] to be suitable for use with adolescents.The 38 items are rated on a 4-point Likert scale withhigher scores indicating a higher frequency with whichthe adolescent has engaged in risk-taking or self-harming behaviors. The RTSHIA has been shown tohave very high reliability (Cronbach’s alpha = 0.89, test–retest reliability = 0.93) and validity.Externalizing and internalizing symptoms are measured

by the Youth Self-Report (YSR) [83] and a correspondingparents’ version, the Child Behavior Checklist (CBCL) [84].One hundred twelve items are rated on a 3-point scalewith “0” for not true, “1” for somewhat or sometimes true,or “2” for very or often true. Both the internalizing and theexternalizing scales comprise symptom scales correspond-ing to DSM diagnoses. A total problem score is derived bysumming up all the symptom scales. The YSR and theCBCL are established evidence-based assessment instru-ments [85], which have also been validated and standard-ized in a Danish sample [86, 87].Borderline personality disorder symptoms are measured

with the Zanarini Rating Scale for Borderline PersonalityDisorder (ZAN-BPD) [88], which is a semi-structuredinterview with ratings from 0 (no symptoms) to 4 (severesymptoms) on each of nine items corresponding to thenine DSM-IV criteria for borderline personality disorder.ZAN-BPD has been used in previous studies of pharmaco-logical and psychological treatments for people with BPD.It is reliable, sensitive to change, and has highly conver-gent validity with structured clinical ratings of BPD [89].Global assessment of functioning will be assessed with

the Children's Global Assessment Scale (C-GAS) thathas shown validity in measuring functioning in children[90] and sufficient inter-rater reliability in a Danishsample [91].Parental reports. In addition to the parents’ version

of the YSR, the CBCL, described above, we will alsocollect parental reports on borderline personality fea-tures using the parents’ version of the BPFS-C, theBPFS-P [78].Sociodemographic information will also be collected at

baseline during assessment procedures and from medicalrecords.

Furthermore, reduction in the number of patients’ hos-pital admissions and visits to the emergency room will beused as a secondary outcome measure. All patients in-cluded in the study will be followed in the National HealthRegister, and we will extract data from the register datingback to the date before the patients were enrolled in thestudy as well.

Mediational variablesAttachment will be assessed with the Experience of CloseRelationships Inventory (ECR) [92] and with the Inventoryof Parent and Peer Attachment-Revised (IPPA-R) [93].The ECR is a 36-item self-report questionnaire measuringattachment in romantic relationships. It has displayedgood psychometric properties. The IPPA-R is a valid andreliable 53-item measure of attachment in adolescence. Itcomprises two scales that measure attachment to parentsand peers, respectively.Mentalization will be assessed with the 46-item measure

Reflective Function Questionnaire for Youth (RFQ-Y) [94].

Measurement intervalsThe primary and all self-report-based secondary outcomesare measured at baseline, week 10, 20, 30, and after the 40th

and final MBT group session. Secondary outcomes basedon interviews and expert ratings (the ZAN-BPD, C-GAS)are measured at baseline and discharge. The test-battery de-signed for intake, including basic clinical evaluation, media-tional and outcome measures will take an estimated time of6 hours per patient. Table 1 lists the primary and secondaryoutcomes, along with mediational measures.We will have weekly self-report-based assessment in

relation to group sessions (40 measurements) and pa-tients will fill out questionnaires in the beginning of eachgroup session. The assessment comprises a short formof the BPFS-C, the BPFS-C-11, and the short version ofthe Youth Self-Report, the BPM-Y (19 items). This is tobe able to monitor closely when in the treatment pro-gram possible progress, and development in the patients’psychopathology is obtained.

Table 1 Assessments administered at baseline and each follow-up point throughout the trial

Assessment points Outcome and mediational self-report measures Expert ratings and clinician-administered measures

Baseline Patient: BPFS-C, YSR, BDI-Y RTSHIA, ECR-R, IPPA-R, RFQ-YParent: CBCL, BPFS-P

Patient: C-GAS, ZAN-BPD

10 weeks, 20 weeks, 30 weeks Patient: BPFS-C, YSR, BDI-Y RTSHIA, ECR-R, IPPA-R, RFQ-YParent: CBCL, BPFS-P

Discharge Patient: BPFS-C, YSR, BDI-Y, RTSHIA, ECR-R, IPPA-R, RFQ-Y.Parent: CBCL, BPFS-P

Patient: C-GAS, ZAN-BPD

BPFS-C Borderline Personality Features Scale for Children, YSR Youth Self-Report, BDI-Y Becks Depression Inventory for Youth, RTSHIA Risk-Taking and Self-HarmInventory for adolescents, ECR-R Experience of Close Relationships Inventory-Revised, IPPA-R Inventory of Parent and Peer Attachment–Revised, RFQ-Y ReflectiveFunction Questionnaire for Youth, C-GAS Children's Global Assessment Scale, ZAN-BPD Zanarini Rating Scale for Borderline Personality Disorder

Beck et al. Trials (2016) 17:314 Page 7 of 13

All applied instruments will be applied in Danishtranslated versions. Published Danish versions exist ofthe SCID-II, the CBCL, the YSR, the BDI-Y, the ECR-Rand the ZAN-BPD. The authors translated the remaininginstruments.

DeteriorationAny patient whose mental health deteriorates duringtreatment will be taken care of according to official guide-lines and treatment recommendations in the Department.We will perform subgroup analyses of patients with anincreased score on the primary outcome at discharge (i.e.,after the 40th session), in order to enhance our knowledgeabout what characterizes patients who do not benefit ordeteriorate from treatment [95].

Sample sizeSample size is determined for the primary outcome,namely, the total score of the BPFS-C. A 12-point differ-ence between the two treatment groups is considered tobe clinically important. The only outcome study usingthe BPFS-C to date reported a standard deviation of out-come of 15.4 for psychotherapeutic treatment with a pa-tient group similar to ours. With 90 % power and a two-tailed significance level of 5 %, 72 patients would needto be randomized to the two treatment arms. This calcu-lation does not, however, take into account the similarityof patients who are treated by the same therapists andpotential dropout.Patients in the MBT group will be treated in five

groups with two therapists in each group. These are slowopen groups with six to seven patients, which will in-clude a new patient as soon as another patient has fin-ished treatment. Group treatments are known to causesimilarity of the outcomes within groups, which needs tobe accounted for in the sample size planning. We expectan intraclass correlation of 0.03, and a total number ofeight patients treated in each of the groups, the designeffect is 1.24. Thus, the sample size has to be increasedby 24 % (which would be 90 patients). For simplicity, weassume the same intraclass correlation in the controlgroup because the control patients are nested withintherapists even if the treatment is individual.Intention-to-treat (ITT) analysis will be used, as it is the

recommended approach to evaluating RCTs ITT analysisincludes every subject who is randomized according to ran-domized treatment assignment and it ignores noncompli-ance, protocol deviations, withdrawal, and anything thathappens after randomization. Hence, ITT analysis preservesprognostic balance produced from the original randomtreatment allocation, but the treatment effect generatedfrom ITT analysis is generally conservative. We will recruit20 % more patients to have powerful tests in both the ITTas well as in the per protocol population even in the

presence of dropout. This yields a total of 112 patients tobe randomized, in a 1:1 ratio.

Randomization, methods to minimize bias, and blindingIn order to keep assessors blind to the patients’ treatmentallocation, all pretreatment assessments will be carried outprior to randomization. Randomization to either MBT orTAU are done online by the trial coordinator (the firstauthor) using a stratified block randomization procedurewith a computer-generated allocation sequence with avarying block size kept unknown to the investigators byPublic Health and Quality Improvement Data Manage-ment. Stratification will be according to clinic affiliationand borderline severity, that is, scores on the BPFS-C.More stratification variables are relevant (i.e., socioeco-nomic status), but not possible to apply due to the smallnumber of participants. Patients randomized to the MBTwill initiate the individual case formulation sessions (seeFig. 1) prior to MBT-I, and thereby will initiate the treat-ment program immediately. Patients randomized to TAUare allocated to a therapist and begin individual sessionsimmediately.Assessments during the treatment phase are limited to

self-report measures, hence no blinding will be possible.All information given to the participants before complet-ing the self-reports will be standardized.Regarding outcome assessments at discharge, we will

minimize bias that knowledge of treatment allocationcould cause by implementing the following strategies: (a)outcome assessors will be blind to treatment allocation,(b) outcome assessors and therapists will not directlycommunicate with each other, and (c) patients are askednot to reveal their treatment allocation during outcomeassessments. Furthermore, outcome assessors will beasked to guess the patients’ treatment allocation so thatthe effects of possible bias can be examined in the analysisand all outcome assessor interviews will be recorded and arandom sample will be re-rated by independent raters. Out-come assessors will be research assistants or clinical psy-chologists, who have not been trained in MBT, have notbeen involved in delivering treatment to participating pa-tients, and have not carried out any intake procedures, in-cluding baseline assessments. The statistician will performstatistical analyses with the two intervention groups codedas ‘A’ and ‘B’, with randomly chosen therapist identifications.

Data management and statistical approachData-management will be handled by Public Health andQuality Improvement Data Management from CentralDenmark Region, Aarhus. They will provide supportwith randomization, set up questionnaires electronicallyand will keep the data on their secure servers. Hence,none of the staff members involved in the study willhave access to outcome data during the treatment phase.

Beck et al. Trials (2016) 17:314 Page 8 of 13

All analyses will be conducted according to the ITTprinciple. Characteristics of the treatment groups willbe described at baseline. Preliminary analysis will in-vestigate the pattern of missingness at follow-up, andmultiple imputation will be used for missing values.The primary outcome is the total score of the BPFS-C, which is treated as a continuous, normally distrib-uted variable. The primary efficacy hypothesis will betested using a multilevel two-group comparison (MBTvs. TAU), with Group as a main effect, TherapyGroup as a random intercept (within the MBT group,see, e.g., [96]), the prerandomization BPFS-C as acontinuous covariate and the stratification factorclinic as categorical covariate. The test will be per-formed at the 5 % two-tailed significance level. Theprimary test for efficacy will be based on theintention-to-treat population, with all randomized pa-tients entering the analysis set, and multiple imput-ation of missing values. The result will be expressedas the covariate-adjusted difference in group averages,along with the two-sided 95 % confidence interval.For the secondary outcomes, similar analyses will be

used, taking into consideration the scale of the variable(e.g., logistic regression for binary outcomes). A sensitiv-ity analysis will be based on the per protocol set for theavailable cases (without imputation). Linear regressionand logistic regression analyses will be conducted toexamine the predictive power of the different covariates(i.e., personality pathology, mentalizing, and attachment)on treatment efficacy and treatment completion.We will use multiple imputations for missing values in

the primary and secondary outcomes [97]. Separate im-putation models will be used for the two treatment groups[98]. The imputation models will include the primary andsecondary endpoints (baseline and follow-up) as well asthe covariates of the primary statistical analysis, and ther-apist time in hours.The primary analysis will be based on the pragmatic

comparison of the two treatment arms and will, there-fore, not take different amounts of therapist time withinor across groups into account. To account for possibledifferences in therapist time, a sensitivity analysis will beperformed on the change scores in which the obtainedvalues will be divided by the actual patient-specific hoursof therapist contact.Mediation analyses will be carried out to test for po-

tential mediating effects of both mentalizing and attach-ment. As recommended by Hayes [99], we will conductstructural equation analyses (SEM), using Preacher andHayes [100] methods for estimating indirect and directeffects with multiple mediators. Unstable participationin the therapy is expected. A dropout will be determinedby clinicians as ending therapy without agreement withthe therapist.

Ethical considerations and regulatory approvalThere are no immediate ethical problems regarding thistrial. Research has not identified any significant adverse ef-fects or risks from any of the compared interventions, andwe have no advance knowledge of which intervention ismost efficacious. During the trial period, any adverse eventwill be reported.The trial is approved by the Regional Ethics Commit-

tee of Zealand (no: SJ-371), and is registered at theDanish Data Protection Agency (no: REG-55-2014).The trial is registered under Clinical Trials.org asNCT02068326. In accordance with the CONSORTguidelines [101, 102], we will report positive, negative,and neutral findings in the trial and we have completedthe SPIRIT 2013 checklist (please see Additional file 1)and figure (please see Additional file 2).

ConfidentialityAll study-related information will be stored securely at thestudy site. All participant information will be stored inlocked file cabinets in areas with limited access. All datawill be identified by a coded identification number only tomaintain participant confidentiality. All records that con-tain names of other personal identifiers such as locatorforms and informed consent forms will be stored separ-ately from study records identified by code number.

Protocol amendmentsAny modifications to the protocol, which may affect theconduct of the study, including changes of study objec-tives, study design, patient population, sample sizes, orstudy procedures, will require a formal amendment tothe protocol. Such amendments will be registered athttps://clinicaltrials.gov/ct2/show/NCT02068326 and ap-proved by the ethics committee.

DiscussionBPD typically emerges in adolescence and the recom-mended primary treatment is psychotherapeutic pro-grams. However, only recently has research begun toinvestigate essential questions such as the relative effect-iveness of separate program components and which levelof specialization versus generalist treatment is neededfor treatment of adolescent BPD to be successful [103].The evidence base for cost-effective treatment programsfor adolescent BPD is limited.This is the first randomized controlled trial to test the

efficacy of a MBT in groups for adolescents with BPD orsubthreshold BPD. Below, we discuss some of the poten-tial limitations and strengths of the M-GAB trial.The trial is subject to at least three potential limita-

tions. First, our primary outcome measure, the BPFS-C,was selected because it reliably and validly measuresborderline personality features specifically in adolescents

Beck et al. Trials (2016) 17:314 Page 9 of 13

(rather than in adults) and has shown to be a sensitiveoutcome measure for adolescents in a previous random-ized trial [53]. However, although patient-reported out-comes have obvious benefits [104], they are not observerblinded and therefore at risk of being assessed withsome bias. In our trial, patients allocated to MBT maybe biased by their knowledge of having received a rela-tively specific psychotherapeutic treatment and thereforebe more optimistic about their outcome. In response tothis potential limitation, we included an observer ratedmeasure of change in BPD symptoms, the ZAN-BPD[88], as a secondary outcome.Second, no standardizations for treatment of adolescent

BPD pre-existed within the child and adolescent depart-ment, Region Zealand and for the purpose of this study,the limit for the minimum treatment in the TAU arm wasdefined as 12 monthly individual sessions. Individual treat-ment allows therapists to focus more intensely on thepatient compared to group-based treatment and therapistsin the TAU arm may choose to deliver treatment to somepatients with a higher frequency than once a month.However, this does not completely rule out the risk of adose–response effect since MBT is delivered on aweekly basis. We defined the minimum frequency forTAU treatment in order to minimize the possible effectof dose–response, but as completely equal doses oftreatment in the two treatment arms is not guaranteed,the risk is not completely ruled out. We will monitorand register both the MBT and TAU patients’ contactto the health system during the treatment period, to beable to report on any differences in the treatment dosesreceived.Third, the MBT therapists are trained and supervised

by engaged and specialized supervisors in a specifictreatment model. It is possible that, as the supervisorsbelieve in the advantages of the MBT model, this mayinfluence the therapist allegiance. We were aware of thepotential allegiance bias when designing the trial, butwanted to ensure therapist adherence to treatment model.The alternative of having less training or less specializedsupervisors increases the risk of therapists not beingadherent to the treatment model, which we considered aneven greater risk to the trial.We would also like to emphasize at least three possible

strengths of the trial. First, we included borderline sever-ity as a stratification variable in order to minimize therisk of confounding our findings by an unbalanced dis-tribution of patients with BPD at the threshold versussubthreshold level. Second, we included outcome mea-sures relying on different methods and sources of data(self-report, parental reports, observer-rated interviews andexpert ratings) and covering different aspects of quality oflife such as symptomatic improvement and social function-ing. By including multiple outcome measures, we hope to

be able to discuss outcome in broader terms. Third, in theselection of in- and exclusion criteria we aimed for partici-pants to be as close to patients with BPD or subthresholdBPD typically referred to treatment at Child and Adoles-cent Psychiatric Department, Region Zealand. For example,patients living with foster parents or other kinds of paren-tal substitutes and patients presenting comorbidity or sub-stance abuse were included. Accordingly, the findings ofthe trial should have a wide generalizability.

Trial statusThe trial is currently in the recruitment phase. The first pa-tient was included and randomized on 24 September 2015and inclusion is expected to be completed in January 2017.

Additional files

Additional file 1: SPIRIT 2013 checklist: recommended items to addressin a clinical trial protocol and related documents*. (DOC 125 kb)

Additional file 2: M-GAB SPIRIT 2013 Figure. (DOC 58 kb)

AbbreviationsBDI-Y, Becks Depression Inventory for Youth; BPD, borderline personality disorder;BPFS-C, Borderline Personality Features Scale for Children; BPFS-C-11, BorderlinePersonality Features Scale for Children-Short version; BPFS-P, Borderline PersonalityFeatures Scale for Children-Parents’ version; CAT, cognitive analytic therapy; CBCL,Childhood Behavior Checklist; C-GAS, Children's Global Assessment Scale; CI-BPD,Childhood Interview for DSM-IV Borderline Personality Disorder; DBT, dialecticalbehavior therapy; DSM-5, Diagnostic and Statistical Manual of Mental Disorders, fifthedition; ECR, Experience of Close Relationships Inventory; IPPA-R, Inventory ofParent and Peer Attachment-Revised; ITT, intention to treat; MINI-KID, MiniInternational Neuropsychiatric Interview for children and adolescents; MBT,mentalization-based treatment; MBT-G, mentalization-based treatment in groups;MBT-I, mentalization-based treatment-introduction; MGT-P, mentalization-basedpsychoeducation therapy for parents of patients; PAI, Personality AssessmentInventory; PD, personality disorder; RCT, randomized controlled trial; RFQ-Y,Reflective Function Questionnaire for Youth; RTSHIA, Risk-Taking and Self-Harm Inventory for adolescents; SCID-II, Structured Clinical Interview forDSM-IV-Axis II; TAU, treatment as usual; YSR, The Youth Self-Report; ZAN-BPD, Zanarini Rating Scale for Borderline Personality Disorder

AcknowledgementsThe authors would like to thank Carla Sharp for her helpful comments andadvice on selection of outcome measures, Trudie Rossouw and Peter Fonagyfor discussing sample size and primary outcome measure with us, AnthonyBateman and Sigmund Karterud for their indispensable advice on the designof the MBT in Groups program, and Michael Maagensen for assisting in thedevelopment of the medical protocol.

Availability of data and materialsAll data will be available for other research groups interested in conductingsystematic reviews. Requests for data should be directed to the PrimaryInvestigator, Professor Erik Simonsen, es@regionsjaelland.dk.

Authors’ contributionsSB conceived the initial idea of a RCT study with MBT in groups for adolescentswith personality pathology. EB and SB developed the trial design withcontributions and supervision from ES, SP, JP, and LP. EB drafted themanuscript, which was carefully edited and discussed by SB, ES, SP, MG, JP,and LP. MG designed the plan for statistical analysis and wrote the sections onsample size and statistical approach. SB modified the MBT-I manual and theMBT-P sessions with assistance from LP. SB and LP ran pilot treatment groupsand derived experiences and information as a knowledge base for furthermodification of the MBT-I manual and the MBT-P sessions. ES developed themedical protocol in collaboration with clinical psychiatrists from the Child and

Beck et al. Trials (2016) 17:314 Page 10 of 13

Adolescent Psychiatric Department Region Zealand. All authors read andapproved the final manuscript.

Competing interestsWe have received external funding for the trial from the Health Science Fund,Region Zealand, Denmark and “Region Sjaellands Forskningsfremmende Pulje”(both are governmental funding). The amount of funding so far is 133.883EUR(salary for co-workers, tuition fee for the university, supervision, etc.).There are no commercial sponsors. The authors declare that they haveno competing interests.

Consent for publicationParticipants will be informed of the trial verbally and in writing, and writteninformed consent will be obtained from at least one of the participants’ parentsbefore inclusion. All trial participants may, upon request, be permitted access tofurther information about the project.

Author details1Child and Adolescent Psychiatric Department, Region Zealand, Smedegade16, 4000 Roskilde, Denmark. 2Psychiatric Research Unit, Region Zealand,Fælledvej 6, 4200 Slagelse, Denmark. 3Department of Psychology, Universityof Copenhagen, Øster Farimagsgade 2A, 1353 Copenhagen K, Denmark.4Institute of Clinical Medicine, Faculty of Health and Medical Sciences,University of Copenhagen, Blegdamsvej 9, 2100 Copenhagen, Denmark.

Received: 19 March 2016 Accepted: 8 June 2016

References1. American Psychiatric Association. Diagnostic and Statistical Manual of

Mental Disorders: DSM-5. Arlington: American Psychiatric Association; 2013.2. Hopwood CJ, Zanarini MC. Borderline personality traits and disorder:

predicting prospective patient functioning. J Consult Clin Psychol. 2010;78(4):585–9.

3. Zanarini MC, Frankenburg FR, Reich DB, Fitzmaurice G. The 10-year courseof psychosocial functioning among patients with borderline personalitydisorder and Axis II comparison subjects. Acta Psychiatr Scand. 2010;122(2):103–9.

4. Black DW, Blum N, Pfohl B, Hale N. Suicidal behavior in borderline personalitydisorder: prevalence, risk factors, prediction, and prevention. J Personal Disord.2004;18(3):226–39.

5. Coid J, Yang M, Tyrer P, Roberts A, Ullrich S. Prevalence and correlates ofpersonality disorder in Great Britain. Br J Psychiatry. 2006;188:423–31.

6. Lenzenweger MF, Lane MC, Loranger AW, Kessler RC. DSM-IV personalitydisorders in the national comorbidity survey replication. Biol Psychiatry.2007;62(6):553–64.

7. Samuels J, Eaton WW, Bienvenu OJ, Brown CH, Costa PT, Nestadt G.Prevalence and correlates of personality disorders in a community sample.Br J Psychiatry. 2002;180:536–42.

8. Torgersen S, Kringlen E, Cramer V. The prevalence of personality disorders ina community sample. Arch Gen Psychiatry. 2001;58(6):590–6.

9. Trull TJ, Jahng S, Tomko RL, Wood PK, Sher KJ. Revised NESARC personalitydisorder diagnoses: gender, prevalence, and comorbidity with substancedependence disorders. J Personal Disord. 2010;24(4):412–26.

10. Fossati A, Maffei C, Bagnato M, Battaglia M, Donati D, Donini M, Fiorili M,Novella L, Prolo F. Patterns of covariation of DSM-IV personality disorders in amixed psychiatric sample. Compr Psychiatry. 2000;41(3):206–15.

11. Zimmerman M, Rothschild L, Chelminski I. The prevalence of DSM-IV personalitydisorders in psychiatric outpatients. Am J Psychiatr. 2005;162(10):1911–8.

12. Korzekwa MI, Dell PF, Links PS, Thabane L, Webb SP. Estimating the prevalenceof borderline personality disorder in psychiatric outpatients using a two-phaseprocedure. Compr Psychiatry. 2008;49(4):380–6.

13. Zanarini MC, Frankenburg FR. Pathways to the development of borderlinepersonality disorder. J Personal Disord. 1997;11(1):93–104.

14. Adam A, Breithaupt-Peters M. Persönlichkeitsentwicklungsstörungen bei kindernund jugendlichen ein integrativer ansatz für die psychotherapeutische undsozialpädagogische praxis. Stuttgart: Kohlhammer; 2010.

15. Freeman A, Reinecke MA. Personality disorders in childhood and adolescence.Hoboken: John Wiley & Sons; 2007.

16. Chanen AM, McCutcheon LK. Prevention and early intervention forborderline personality disorder: current status and recent evidence. Br JPsychiatry. 2013;202:24–9.

17. Bondurant H, Greenfield B, Tse SM. Construct validity of the adolescentborderline personality disorder: a review. Can Child Adolesc Psychiatr Rev.2004;13(3):53–7.

18. Chanen AM, Kaess M. Developmental pathways to borderline personalitydisorder. Curr Psychiatry Rep. 2012;14(1):45–53.

19. Miller AL, Muehlenkamp JJ, Jacobson CM. Fact or fiction: diagnosingborderline personality disorder in adolescents. Clin Psychol Rev. 2013;28(6):969–81.

20. Paris J. Personality disorders begin in adolescence. J Can Acad Child AdolescPsychiatry. 2013;22(3):195–6.

21. Sharp C, Romero C. Borderline personality disorder: a comparison betweenchildren and adults. Bull Menn Clin. 2007;71(2):85–114.

22. Glenn CR, Klonsky ED. Reliability and validity of borderline personalitydisorder in hospitalized adolescents. J Can Acad Child Adolesc Psychiatry.2013;22(3):206–11.

23. Grilo CM, McGlashan TH, Quinlan DM, Walker ML, Greenfeld D, Edell WS.Frequency of personality disorders in two age cohorts of psychiatricinpatients. Am J Psychiatry. 1998;155(1):140–2.

24. Becker DF, Grilo CM, Edell WS, McGlashan TH. Diagnostic efficiency ofborderline personality disorder criteria in hospitalized adolescents:comparison with hospitalized adults. Am J Psychiatry. 2002;159(12):2042–7.

25. Cohen P, Crawford TN, Johnson JG, Kasen S. The children in the communitystudy of developmental course of personality disorder. J Personal Disord.2005;19(5):466–86.

26. Cohen P. Child development and personality disorder. Psychiatr Clin N Am.2008;31(3):477–93.

27. Keuroghlian AS, Frankenburg FR, Zanarini MC. The relationship of chronicmedical illnesses, poor health-related lifestyle choices, and health careutilization to recovery status in borderline patients over a decade ofprospective follow-up. J Psychiatr Res. 2013;47(10):1499–506.

28. Ha C, Balderas JC, Zanarini MC, Oldham J, Sharp C. Psychiatric comorbidityin hospitalized adolescents with borderline personality disorder. J ClinPsychiatry. 2014;75(5):457–64.

29. Zanarini MC, Frankenburg FR, Dubo ED, et al. Axis I comorbidity ofborderline personality disorder. Am J Psychiatry. 1998;155(12):1733–94.

30. McGlashan TH, Grilo CM, Skodol AE, et al. The collaborative longitudinalpersonality disorders study: baseline Axis I/II and II/II diagnostic co-occurrence. Acta Psychiatr Scand. 2000;102(4):256–64.

31. Skodol AE, Gunderson JG, Pfohl B, et al. The borderline diagnosis I:psychopathology, comorbidity, and personality structure. Biol Psychiatry.2002;51(12):936–50.

32. Hörz S, Zanarini MC, Frankenburg FR, Reich DB, Fitzmaurice G. Ten-yearuse of mental health services by patients with borderline personalitydisorder and with other Axis II disorders. Psychiatr Serv (Washington, DC).2010;61(6):612–6.

33. van Asselt AD, Dirksen CD, Arntz A, Severens JL. The cost of borderlinepersonality disorder: societal cost of illness in BPD patients. Eur Psychiatry.2007;22(6):354–61.

34. Cailhol L, Jeannot M, Rodgers R, Guelfi JD, Perez-Diaz F, Pham-Scottez A,Speranza M. Borderline personality disorder and mental healthcare serviceuse among adolescents. J Personal Disord. 2013;27(2):252–9.

35. Feenstra DJ, Hutsebaut J, Laurenssen EM, Verheul R, Busschbach JJ, Soeteman DI.The burden of disease among adolescents with personality pathology: quality oflife and costs. J Personal Disord. 2012;26(4):593–604.

36. Chanen AM, Jovev M, Jackson HJ. Adaptive functioning and psychiatricsymptoms in adolescents with borderline personality disorder. J ClinPsychiatry. 2007;68(2):297–306.

37. Chanen AM, Jovev M, McCutcheon L, Jackson HJ, McGorry PD. Borderlinepersonality disorder in young people and the prospects for prevention andearly intervention. Curr Psychiatry Rev. 2008;4(1):48–57.

38. Winograd G, Cohen P, Chen H. Adolescent borderline symptoms in thecommunity: prognosis for functioning over 20 years. J Child PsycholPsychiatry Allied Discip. 2008;49(9):933–41.

39. Lenzenweger MF, Desantis CD. Predicting change in borderline personality:using neurobehavioral systems indicators within an individual growth curveframework. Dev Psychopathol. 2005;17(4):1207–37.

40. Bleiberg E, Rossouw T, Fonagy P. Adolescent breakdown and emergingborderline personality disorder. In: Bateman A, Fonagy P, editors. Handbook

Beck et al. Trials (2016) 17:314 Page 11 of 13

of mentalizing in mental health practice. Washington DC: AmericanPsychiatric Publishing; 2012. p. 463–509.

41. Bateman AW, Gunderson J, Mulder R. Treatment of personality disorder.Lancet. 2015;385:735–43.

42. Stoffers J, Völlm BA, Rücker G, Timmer A, Huband N, Lieb K. Pharmacologicalinterventions for borderline personality disorder. Cochrane Database Syst Rev.2010;16:CD005653.

43. Mehlum L, Tørmoen AJ, Ramberg M, Haga E, Diep LM, Laberg S, et al.Dialectical behavior therapy for adolescents with repeated suicidal and self-harming behavior: a randomized trial. J Am Acad Child Adolesc Psychiatry.2014;53:1082–91.

44. Mehlum L, Ramberg M, Tørmoen AJ, Haga E, Diep BHS, et al. Dialecticalbehavior therapy compared with enhanced usual care for adolescents withrepeated suicidal and self-harming behavior: outcomes over a one-yearfollow-up. J Am Acad Child Adolesc Psychiatry. 2016;55(4):295–300.

45. Rathus JH, Miller AL. Dialectical behavior therapy adapted for suicidaladolescents. Suicide Life Threat Behav. 2002;32(2):146–57.

46. Woodberry KA, Popenoe EJ. Implementing dialectical behavior therapy withadolescents and their families in a community outpatient clinic. CognBehav Pract. 2008;15:277–86.

47. Fleischhaker C, Böhme R, Sixt B, Brück C, Schneider C, Schulz E. DialecticalBehavioral Therapy for Adolescents (DBT-A): a clinical trial for patients withsuicidal and self-injurious behavior and borderline symptoms with a one-year follow-up. Child Adolesc Psychiatr Mental Health. 2011;5(1):3.

48. Tørmoen AJ, Grøholt B, Haga E, Brager-Larsen A, Miller A, Walby F, Mehlum L.Feasibility of dialectical behavior therapy with suicidal and self-harmingadolescents with multi-problems: training, adherence, and retention. ArchSuicide Res. 2014;18:432–44.

49. James S, Freeman K, Mayo D, Riggs M, Morgan J, Schaepper MA, et al. Doesinsurance matter? Implementing dialectical behavior therapy with twogroups of youth engaged in deliberate self-harm. Adm Policy Ment HealthMent Health Serv Res. 2014;42:449–61.

50. Freeman KR, James S, Klein KP, Mayo D, Montgomery S. Outpatientdialectical behavior therapy for adolescents engaged in deliberate self-harm: conceptual and methodological considerations. Child Adolesc SocWork J. 2016;33(2):123–35.

51. Chanen AM, Jackson HJ, McCutcheon LK, Jovev M, Dudgeon P, Yuen HP,McGorry PD. Early intervention for adolescents with borderline personalitydisorder using cognitive analytic therapy: randomised controlled trial. Br JPsychiatry. 2008;193(6):477–84.

52. Schuppert HM, Timmerman ME, Bloo J, van Gemert TG, Wiersema HM,Minderaa RB, Nauta MH. Emotion regulation training for adolescents withborderline personality disorder traits: a randomized controlled trial. J AmAcad Child Adolesc Psychiatry. 2012;51(12):1314–23.

53. Rossouw TI, Fonagy P. Mentalization-based treatment for self-harm inadolescents: a randomized controlled trial. J Am Acad Child AdolescPsychiatry. 2012;51(12):1304–13.

54. Laurenssen EM, Feenstra DJ, Busschbach JJ, Hutsebaut J, Bales DL, Luyten P,Noom MJ. Feasibility of mentalization-based treatment for adolescents withborderline symptoms: a pilot study. Psychotherapy. 2014;51(1):159–66.

55. Bateman A, Fonagy P. Effectiveness of partial hospitalization in thetreatment of borderline personality disorder: a randomized controlled trial.Am J Psychiatry. 1999;156(10):1563–9.

56. Bateman A, Fonagy P. Treatment of borderline personality disorder withpsychoanalytically oriented partial hospitalization: an 18-month follow-up.Am J Psychiatry. 2001;158(1):36–42.

57. Bateman A, Fonagy P. 8-year follow-up of patients treated for borderlinepersonality disorder: mentalization-based treatment versus treatment asusual. Am J Psychiatry. 2008;165(5):631–8.

58. Bateman A, Fonagy P. Randomized controlled trial of outpatientmentalization-based treatment versus structured clinical management forborderline personality disorder. Am J Psychiatry. 2009;166(12):1355–64.

59. Bateman A, Fonagy P. Psychotherapy for borderline personality disorder:mentalization-based treatment. Oxford. New York: Oxford University Press;2004.

60. Fonagy P, Gergely G, Jurist E, Target M. Affect regulation, mentalization, andthe development of the self. New York: Other Press; 2002.

61. Fonagy P, Bateman AW. Adversity, attachment, and mentalizing. ComprPsychiatry. 2016;64:59–66.

62. Karterud S. Manual for mentaliseringsbasert gruppeterapi (MBT-G). Oslo:Gyldendal akademisk; 2012.

63. Gunderson JG, Zanarini MC, Choi-Kain LW, Mitchell KS, Jang KL, Hudson JI.Family study of borderline personality disorder and its sectors ofpsychopathology. Arch Gen Psychiatry. 2011;68(7):753–62.

64. Carlson EA, Egeland B, Sroufe LA. A prospective investigation of thedevelopment of borderline personality symptoms. Dev Psychopathol. 2009;21:1311–1334.

65. De Panfilis C, Marchesi C, Cabrino C, Monici A, Politi V, Rossi M, Maggini C.Patient factors predicting early dropout from psychiatric outpatient care forborderline personality disorder. Psychiatry Res. 2012;200(2-3):422–9.

66. Karterud S, Bateman AB. Manual for mentaliseringsbasert psykoedukativgruppeterapi (MBT-I). Oslo: Gyldendal akademisk; 2011.

67. Karterud S. On structure and leadership in mentalization-based group therapyand group analysis’. Group Analysis. 2015;48(2):137–49.

68. Sundhedstyrelsen. National Klinisk retningslinje for behandling af emotionelustabil personlighedsstruktur, borderline type. Kørbenhavn:Sundhedsstyrelsen; 2015.

69. Sheehan DV, Lecrubier Y, Sheehan KH, Amorim P, Janavs J, Weiller E,Hergueta T, Baker R, Dunbar GC. The Mini-International NeuropsychiatricInterview (M.I.N.I.): the development and validation of a structureddiagnostic psychiatric interview for DSM-IV and ICD-10. J Clin Psychiatry.1998;59(20):22–33.

70. Sheehan DV, Sheehan KH, Shytle RD, Janavs J, Bannon Y, Rogers JE, et al.Reliability and validity of the Mini International Neuropsychiatric Interviewfor Children and Adolescents (MINI-KID). J Clin Psychiatry. 2010;71(3):313–26.

71. First M, Gibbon M, Spitzer R. Structured clinical interviews for DSM-IV Axis-iipersonality disorders. Washington DC: American Psychiatric Press; 1997.

72. Lobbestael J, Leurgans M, Arntz A. Inter-rater reliability of the structuredclinical interview for DSM-IV Axis I disorders (SCID I) and Axis II disorders(SCID II). Clin Psychol Psychother. 2011;18(1):75–9.

73. Zanarini MC. The childhood interview for DSM-IV borderline personality disorder(CI-BPD). Belmont: McLean Hospital and Harvard Medical School; 2003.

74. Sharp C, Ha C, Michonski J, Venta A, Carbone C. Borderline personalitydisorder in adolescents: evidence in support of the Childhood Interview forDSM-IV Borderline Personality Disorder in a sample of adolescent inpatients.Compr Psychiatry. 2012;53:765–74.

75. Sharp C, Ha C, Carbone C, Kim S, Perry K, Williams L, Fonagy, P. Hypermentalizingin adolescent inpatients: treatment effects and association with borderline traits.J Personal Disord. 2013;27(1):3–18.

76. Crick NR, Murray-Close D, Woods K. Borderline personality features in childhood:a short-term longitudinal study. Dev Psychopathol. 2005;17(4):1051–70.

77. Morey LC. Personality assessment inventory: professional manual. Odessa:Psychological Assessment Resources; 1997.

78. Sharp C, Mosko O, Chang B, Ha C. The cross-informant concordance andconcurrent validity of the Borderline Personality Features Scale for Childrenin a community sample of boys. Clin Child Psychol Psychiatry. 2010;16(3):335–49.

79. Sharp C, Steinberg L, Temple J, Newlin E. An 11-item measure to assessborderline traits in adolescents: refinement of the BPFSC using IRT. PersonalDisord. 2014;5(1):70–8.

80. Thastum M, Ravn K, Sommer S, Trillingsgaard A. Reliability, validity andnormative data for the Danish Beck Youth Inventories. Scand J Psychol.2009;50(1):47–54. 52.

81. Vrouva I, Fonagy P, Fearon PR, Roussow T. The risk-taking and self-harminventory for adolescents: development and psychometric evaluation.Psychol Assess. 2010;22(4):852-865.

82. Sansone RA, Wiederman MW, Sansone LA. The Self-Harm Inventory:development of a scale for identifying self-destructive behaviors andborderline personality disorder. J Clin Psychol. 1998;54:973–83.

83. Achenbach TM. Manual for the Youth Self-Report and 1991 Profile.Burlington: University of Vermont Department of Psychiatry; 1991.

84. Achenbach TM. Manual for the Child Behavior Checklist/4-18 and 1991Profile. Burlington: University of Vermont Department of Psychiatry; 1991.

85. Holmbeck GN, Thill AW, Bachanas P, Garber J, Miller KB, Abad M, et al.Evidence-based assessment in pediatric psychology: measures of psychosocialadjustment and psychopathology. J Pediatr Psychol. 2008;33:958–80.

86. Bilenberg N. The Child Behavior Checklist (CBCL) and related material:standardization and validation in Danish population based and clinicallybased samples. Acta Psychiatr Scand Suppl. 1999;398:2–52.

87. Henriksen H, Nielsen PF, Bielenberg N. Ny dansk standardisering afadfærdsspørgeskemaet Child Behaviour Checklist til skolebørn. UgeskrLaeger. 2012;174(46):2867.

Beck et al. Trials (2016) 17:314 Page 12 of 13

88. Zanarini MC, Frankenburg FR. Zanarini Rating Scale for Borderline PersonalityDisorder (ZAN-BPD). Belmont: Harvard Medical School; 2001.

89. Zanarini MC, Vujanovic AA, Parachini EA, Boulanger JL, Frankenburg FR,Hennen J. Zanarini Rating Scale for Borderline Personality Disorder (ZAN-BPD):a continuous measure of DSM-IV borderline psychopathology. J Pers Disord.2003;17(3):233–42.

90. Shaffer D, Gould MS, Brasic J, Ambrosini P, Fisher P, Bird H, Aluwahlia S. Achildren’s global assessment scale (CGAS). Arch Gen Psychiatry. 1983;40(11):1228–31.

91. Dyrborg J, Larsen FW, Nielsen S, Byman J, Nielsen BB, Gautrè-Delay F. Thechildren’s global assessment scale (CGAS) and global assessment ofpsychosocial disability (GAPD) in clinical practice–substance and reliability asjudged by intraclass correlations. Eur Child Adolesc Psychiatry. 2000;9(3):195–201.

92. Brennan KA, Clark CL, Shaver PR. Self-report measurement of adult attachment:An integrative overview. In: Simpson JA, Rholses WS, editors. Attachmenttheory and close relationships. New York: Guilford Press; 1998. p. 46–76.

93. Gullone E, Robinson K. The inventory of parent and peer attachment -revised (IPPA-R) for children: A psychometric investigation. Clin PsycholPsychother. 2005;12(1):67–79.

94. Ha C, Sharp K, Ensink P, Fonagy PC. The measurement of reflective function inadolescents with and without borderline traits. J Adolesc. 2013;36:1215–23.

95. Lambert MJ. Outcome in psychotherapy: the past and important advances.Psychotherapy (Chicago, Ill). 2013;50(1):42–51.

96. Baldwin SA, Bauer DJ, Stice E, Rohde P. Evaluating models for partiallyclustered designs. Psychol Methods. 2011;16:149–65.

97. Hayes AF. Beyond Baron and Kenny: statistical mediation analysis in thenew millennium. Commun Monogr. 2009;76(4):408–20.

98. Van Buuren S, Groothuis-Oudshoorn K. mice: Multivariate imputation bychained equations in R. J Stat Softw. 2011;45(3):1–67.

99. Bell ML, Fairclough DL. Practical and statistical issues in missing data forlongitudinal patient-reported outcomes. Stat Methods Med Res. 2014;23(5):440–59.

100. Preacher KJ, Hayes AF. Asymptotic and resampling strategies for assessingand comparing indirect effects in multiple mediator models. Behav ResMethods. 2008;40(3):879–91.

101. Chan A-W, Tetzlaff JM, Altman DG, Laupacis A, Gøtzsche PC, Krleža-Jerić K,Hróbjartsson A, Mann H, Dickersin K, Berlin J, Doré C, Parulekar W, Summerskill W,Groves T, Schulz K, Sox H, Rockhold FW, Rennie D, Moher D. SPIRIT 2013Statement: defining standard protocol items for clinical trials. Ann Intern Med.2013;158:200–7.

102. Chan A-W, Tetzlaff JM, Gøtzsche PC, Altman DG, Mann H, Berlin J, Dickersin K,Hróbjartsson A, Schulz KF, Parulekar WR, Krleža-Jerić K, Laupacis A, Moher D.SPIRIT 2013 explanation and elaboration: guidance for protocols of clinicaltrials. BMJ. 2013;346:e7586.

103. Chanen A, Jackson H, Cotton SM, Gleeson J, Davey CG, Betts J, McCutcheon L.Comparing three forms of early intervention for youth with borderlinepersonality disorder (the MOBY study): study protocol for a randomisedcontrolled trial. Trials. 2015;16:476.

104. Johnston BC, Patrick DL, Busse JW, Schünemann HJ, Agarwal A, Guyatt GH.Patient-reported outcomes in meta-analyses–part 1: assessing risk of biasand combining outcomes. Health Qual Life Outcomes. 2013;11:109.

• We accept pre-submission inquiries

• Our selector tool helps you to find the most relevant journal

• We provide round the clock customer support

• Convenient online submission

• Thorough peer review

• Inclusion in PubMed and all major indexing services

• Maximum visibility for your research

Submit your manuscript atwww.biomedcentral.com/submit

Submit your next manuscript to BioMed Central and we will help you at every step:

Beck et al. Trials (2016) 17:314 Page 13 of 13